Muscle Fatigue Estimation with Wearable EMG Sensors: A Case Study of Lifting Dumbbell

• Main Authors: Kuan-Chun Lin, 林冠錞
• Other Authors: Ming-Sui Lee
• Format: Others
• Language: en_US
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/7rs6z5

碩士 === 國立臺灣大學 === 資訊工程學研究所 === 104 === Due to the smartphones and SoC are more and more convenient and small, wearable devices become common and popular recently. People start to use wearable devices to record their body status. Heart rate sensors, pedometers are seen in smartwatches in large quantities these two years. Also, in the future, EMG sensor will very likely play an important role in wearable devices. A new type of input device which utilize EMG sensor came out several months ago. Furthermore, fitness is a very popular high intensity activity. However, high intensity activity comes with high risk injury. As a result, utilizing EMG sensor and the trend to recording users’ own information, we developed a Arduino system which combine EMG sensor and feedback of muscle fatigue level when fitness. In this thesis, we proposed a system for estimating users’ level of muscle fatigue. At the beginning, we separate whole fitness activity into “action” from raw signal. After then, the signal is performed by Fourier transform and unwanted frequencies are eliminated. Finally, we normalize the total energy, which is generated by motor unit, and apply to our model to estimate the level of users’ muscle fatigue. In experiment, a lifting dumbbell trial was designed as the exhausting exercises. In summary, we proposed a method to separate each action during lifting dumbbell clearly and a new aspect of muscle fatigue evaluation; and results show that our system is able to estimate the level of muscle fatigue with 10% errors between evaluation and reality.